Comparison of Four Precipitation Based Meteorological Drought Indices in Yesilirmak Basin, Turkey

2 Drought, which is often defined as not enough precipitation, does not a mean simple lack of precipitation. This 3 condition, which occurs when humidity is less than the average value for many years, is caused by a disrupted balance 4 between precipitation and evaporation in a region. It is very difficult to predict the start and the end time of drought. 5 In the present study, the drought conditions of the stations selected from Yesilirmak Basin between 1970 and 2014 6 were determined by using Z-Score Index (ZSI), China-Z Index (CZI), Modified China-Z Index (MCZI), and Standard 7 Precipitation Index (SPI), and the compliance of these indices to the SPI was investigated. It was determined that these 8 indices gave parallel results to each other, and SPI detected drought earlier than other indices.


Introduction
which is a semi-arid area in Iran, and analyzed SPI, RDDI and CZI relatively to each other. They found that SPI yielded 1 the best results after using 41 years' average precipitation data from eight different precipitation stations.

2
In this study, the purpose was to calculate, compare, and evaluate four meteorological drought indices (ZSI, 3 CZI, MCZI, SPI) for various time periods to identify droughts in Yesilirmak Basin, and also to examine the relations 4 between drought indices and time periods. The total monthly precipitation values of Amasya, Çorum, Samsun and 5 Tokat meteorological observation stations located in Yesilirmak Basin were used. When longer records are used to 6 calculate drought indices, more reliable results can be obtained (Wu et al. 2001). For this reason, applications were 7 made for a 45-year period between 1970 and 2014, which was the longest data range for the specified meteorology  Table 1 12 (2)

14
In Eqs. (1) and (2), x refers to the amount of precipitation, and Γ(α) is the Gamma Function. SPI requires that 15 a Gamma probability density function is adapted to frequency distribution given with precipitation totals for a station.

16
The shape (α) and scale (β) parameters of the gamma probability density function are predicted for each station and 17 time period in question. The maximum probability solutions given by Thom (1958)     (4) The gamma distribution is non-defined for zero values of x; however, since the precipitation series may 1 contain zero values, the cumulative probability distribution for zero precipitation and precipitations other than zero is 2 identified as equation (6)  3 4 In equation (6), "q" is the probability of zero. If "m" is the number of zeros in the precipitation time series, it 5 can be predicted as q=m/n. The probability function H(x) is converted into SPI that has an average of zero and a    Table 2.

18
Raw precipitation data are used in ZSI method, which is a unidimensional drought index. As seen in Eq. (9),

24
In Eq. (9), x i refers to the precipitation values in the time period, x ̅ refers to the average precipitation data,

25
and σ refers to the Standard Deviation. The drought classification according to ZSI is given in

14
In Eq. (11), φ j is the standard variable, and Me refers to the median value of precipitation. The drought 15 classification according to MCZI value is given in Table 2.            and 24 months for SPI-12; and 24 months, 28 months, 40 months, and 55 months for SPI-24.

3
It was seen that SPI determined drought earlier than other indices used. This was evident in Samsun and Tokat 4 for 12-months index values in all stations for 24-months index values. It was seen that the fact that the SPI determined 5 the drought earlier was a remarkable feature of the index.

6
The correlation matrix and scatter diagrams of the stations were also prepared to examine the agreement better

1
In the present study, drought analysis was made for Yesilirmak Basin, which is one of the basins of Turkey 2 with water potential and drought risk. The data for 4 meteorological stations selected from the basin between 1970 and 3 2014 were obtained from the Turkish State Meteorological Service. Four different meteorological drought indices 4 (ZSI, CZI, MCZI and SPI), which required precipitation data were calculated in three-time scales (3-months, 12-5 months, and 24-months); and drought quantities (intensity, duration) were examined. Also, the relation of the indices 6 with SPI, which was selected as the reference index, was investigated and evaluated.